Vulnerabilities
AI Coding Agent Skills Bypass SCA and SAST Security Scanners
Research reveals that malicious instructions in AI agent skills evade static code analysis, creating blind spots for credential theft and data exfiltration in development workflows.
